Databases Reference
In-Depth Information
that knowledge and form conceptual and logical schemas, how these are
schemas validated and transformed into low-level representations, and how
they are verified and validated. Our aim is to provide the basic ideas that gov-
ern the design and implementation of a CASE tool and to show the balance
between what a CASE tool can do and what remains the designers creativity
and decisions. We particularly insist in these sections on CASE functionali-
ties that help in solving hard problems, such as knowledge acquisition, con-
ceptual modeling, and design validation.
13.2
A CASE Framework for Database Design
Database design has been widely investigated and explored during the past
three decades. Many design frameworks have been proposed, and there is a
consensus to distinguish among four abstraction levels: external, conceptual,
logical, and physical design. Based on these levels, different modeling nota-
tions, techniques, and approaches have been proposed. Early provided design
tools support relational normalization, schema mapping between the entity-
relationship model and the relational model, and DDL generation. The early
1980s saw the promotion of expert systems and knowledge-based tools that
integrated heuristics, design alternatives, and high-level interaction with
the human designer [6]. The late 1980s confirmed the industrial use of DB
design tools; hundreds of CASE tools were proposed in the software engi-
neering market. The 1990s saw the emergence of object-oriented languages
and methodologies with their companion tools. Database design tools gained
in maturity and in complexity.
To understand the role and the contribution of these tools, we use the
framework in Figure 13.3. The framework serves as an ideal CASE environ-
ment, one that illustrates most of the possible tools related to DB design.
Knowledge acquisition concerns the collection of all the knowledge
necessary for the conceptual modeling of the DB. Knowledge acquisition is
done during user requirements analysis, either by interaction with potential
DB users, extraction of data from forms and texts, or by the use of some
appropriate graphical interface. Knowledge acquisition is driven by preexist-
ing domain knowledge, a predefined enterprise model, or any procedure that
helps in requirement analysis.
Data abstraction and structuring consist of organizing the knowledge
acquired during the acquisition phase and defining the main entities and
relationships that best capture the views of the users. That corresponds to
the effective conceptual modeling phase. Depending on the complexity of
